The update script in the VM: updates-docs-full.sh sometimes hangs and it doesn't let further updates to kick in. This is because the script locks a file while it is running, and cron trying to update docs while the previous run didn't complete would not go through. In such an instance, I had to manually kill the hung script & lock file:  https://crbug.com/682376 

1) I do not have any idea why the script hangs. There might be a genuine bug in update_cache.py. Investigate what might be going on here.
2) We might want to considering timeouts: if a previous run of the script didn't finish in some given time, then give up.
3) Using the file lock (flock) is great to ensure we don't try to write to the datastore file while another one is in-flight, but it might not be the best way... Investigate some other options.